[washingtonismoney]

And, more importantly, we should have attacked Japan before Perl Harbor. It would appear you are saying we should make all important decisions (correctly, naturally) before the season starts and ignore what happens during the season. I doubt you meant that. It would be nice if that were possible, but maybe it isn't?

No, I'm making a much different point. Before Japan attacked Pearl Harbor, there was no casus belli to make war with Japan for the United States. Hence the actual event of Pearl Harbor changed everything about the U.S.'s position in the world.

Player performance -- and what level of performance justifies a change -- is much subtler. As we know there are issues of assessment (do we have a sufficient sample size to judge whether a given player is good? will a player improve? what do we know about his replacement, etc.).

So the question is: "what level of information is sufficient to know it's worth switching?" What my comment is suggesting is that it takes a great deal of valuable information to justify a switch. Since that's the case, if it's worth making the switch in game 3, it was probably also worth making the switch before the season began.

That's not always the case. For example, a serious injury would automatically justify a switch. A truly horrendous performance might also justify the switch. But that wasn't the case; we'd seen one game with reasonably effective performance, albeit against weak opposition, and one game in which the offensive line seemingly performed poorly. But even then there was a great deal of uncertainty: was it just an SC thing? (SC really got after Texas, so ... maybe?)  was it something distinctive about the scheme or matchup that made the offensive line (seemingly) ineffective? (SC seemed devoted to crowding the line and blitzing relentlessly. If you bring tons of rushers the line won't inherently look good. But theoretically you can bust that defense with big plays -- which is more-or-less what Stanford was able to do against SC.) And so on.

So it wasn't, at least for outside observers, conclusive at all. Of course coaches have access to information we don't. That means, often, we should give the benefit of the doubt, particularly if coaches have strong track records. But we shouldn't always give that benefit of the doubt. Sometimes coaches with strong track records do stuff like, "bring in a QB specifically to serve as a backside blocker of a play." Then you question them. That's what makes a remark like mine sensible. Which I think it is, even if it isn't strictly speaking right.

[washingtonismoney]

And, more importantly, we should have attacked Japan before Perl Harbor. It would appear you are saying we should make all important decisions (correctly, naturally) before the season starts and ignore what happens during the season. I doubt you meant that. It would be nice if that were possible, but maybe it isn't?

No, I'm making a much different point. Before Japan attacked Pearl Harbor, there was no casus belli to make war with Japan for the United States. Hence the actual event of Pearl Harbor changed everything about the U.S.'s position in the world.

Player performance -- and what level of performance justifies a change -- is much subtler. As we know there are issues of assessment (do we have a sufficient sample size to judge whether a given player is good? will a player improve? what do we know about his replacement, etc.).

So the question is: "what level of information is sufficient to know it's worth switching?" What my comment is suggesting is that it takes a great deal of valuable information to justify a switch. Since that's the case, if it's worth making the switch in game 3, it was probably also worth making the switch before the season began.

That's not always the case. For example, a serious injury would automatically justify a switch. A truly horrendous performance might also justify the switch. But that wasn't the case; we'd seen one game with reasonably effective performance, albeit against weak opposition, and one game in which the offensive line seemingly performed poorly. But even then there was a great deal of uncertainty: was it just an SC thing? (SC really got after Texas, so ... maybe?)  was it something distinctive about the scheme or matchup that made the offensive line (seemingly) ineffective? (SC seemed devoted to crowding the line and blitzing relentlessly. If you bring tons of rushers the line won't inherently look good. But theoretically you can bust that defense with big plays -- which is more-or-less what Stanford was able to do against SC.) And so on.

So it wasn't, at least for outside observers, conclusive at all. Of course coaches have access to information we don't. That means, often, we should give the benefit of the doubt, particularly if coaches have strong track records. But we shouldn't always give that benefit of the doubt. Sometimes coaches with strong track records do stuff like, "bring in a QB specifically to serve as a backside blocker of a play." Then you question them. That's what makes a remark like mine sensible. Which I think it is, even if it isn't strictly speaking right.